Skip to contentSkip to conversation

Implement the NVMe reservation feature

PATCH Superseded Guixin Liu 6 participants 19 messages Oct 21, 2024 – Oct 24, 2024

Subsystems (2)

v17 of 19 Superseded Latest is v19 →

lore
Share & tools
lore

Guixin Liu

author
#0 lore raw

v17 cover Implement the NVMe reservation feature

Hi guys,
I've implemented the NVMe reservation feature. Please review it, all
comments are welcome as usual.

Changes from v16 to v17:
- Put nvmet_pr_do_abort() into workqueue to not block the the transport thread.
- Fix a bug in nvmet_pr_get_ns_pc_ref(), we should set req->pc_ref after percpu_ref_tryget_live() scucess, otherwise the pc_ref may be decremented one extra.

Changes from v15 to v16:
- Fix kerenl message print "INFO: trying to register non-static key", when the namespace is not enable reservation.
- Fix the race condition between nvmet_ctrl_init_pr() and nvmet_ns_enable() by putting nvmet_ctrl_init_pr() and nvmet_ctrl_destroy_pr() into the brace of mutex_lock(&subsys->lock). Avoid the double init of per-controller percpu ref.

Changes from v14 to v15:
- Change first patch commit message.

Changes from v13 to v14:
- Fix some grammatical mistake in commit message.
- Add two characters to bulleted lists in commit message to make it more readble.
- Adjust some tab indent.

Changes from v12 to v13:
- Set regctl to the number of all registrants instead of reported when reporting.

Changes from v11 to v12:
- Change patch to fit newest 6.12-rc2.
- Fix possible out-of-bounds access in nvmet_execute_pr_report.
- Use sysfs_emit instead of printf in nvmet_ns_resv_enable_show.
- Add struct rcu_head to nvmet_pr_registrant to free.
- Introduce a helper nvmet_pr_parse_ignore_key.
- Use NVME_CNTLID_DYNAMIC when report cntlid.
- Fix some over long line.
- Add more details to commit body.
- Separate out the include/linux/nvme.h changes to a single patch.
- Add a nvmetcli patch to support resv_enable configuration.

Changes from v10 to v11:
- Remove the cntlid from struct nvmet_pr_registrant. And add hostid to nvmet_pr_per_ctrl_ref to set the correct ctrl to do abort.
- Report all registrants, and report cntlid 0xffff.
- Change nvmet_req's flags to pc_ref.

Changes from v9 to v10:
- Fix the misjudgement in nvmet_pr_unregister_one.
- Fix the non-atomicity problem in nvmet_pr_preempt, now we set the new holder first before unregistering, this can not only make sure that during unregistering other host can not access, but also ensure that nvmet_pr_unregister_one will not unregiter the new holder(In nvmet_pr_unreg_all_others_by_prkey, I exclude current host).
- Remove the pr_abort in nvmet_ctrl, instead kill the per-controller percpu ref first and then wait to zero at the end of preempt_and_abort.
- Fix some spelling mistakes.

Changes from v8 to v9:
- Remove the maintianer request.

- Support "preemt and abort" by adding a per-controller percpu ref to ns, doing wait per-controller percpu ref to zero when a controller's reservation or registration is preempt and the racqa == "preempt and abort". Currently add per-controller percpu ref only when nvmet_pr_check_cmd_access success, others are not cared.

- Report ns support reservation when resv is enabled.

- Report ctrl support reservation.

- Change the log level to info when log lost.

- Fix the UAF issue in nvmet_pr_unreg_by_prkey, and change nvmet_pr_unreg_by_prkey to nvmet_pr_unreg_all_host_by_prkey.

- Dont unregister the host when ctrl is destroyed for keep the reservation info when reconnect.

- Remove the rcu lock and mutex lock when free ns's pr info.

- Fix the situation of log.count is zero.

- Move the rtype check to the start of preemtion, Dmitry suggests that:
1.4.1.6 reserved
Receipt of reserved coded values in defined fields in commands shall be reported as an error.
Look forward the suggestions.
And also, this avoid the non-atomic change of unregister and set new holder.

- Fix the compile error when close CONFIG_LOCKDEP.

- Change nvmet_pr_send_event_by_hostid to nvmet_pr_send_event_to_host.

- Change nvmet_pr_unreg_by_prkey_except_hostid to nvmet_pr_unreg_all_others_by_prkey.

- Change nvmet_pr_unreg_by_prkey to nvmet_pr_unreg_all_host_by_prkey.

Changes from v7 to v8:
- Add me as the new file pr.c's maintainer.

Changes from v6 to v7:
- Handle "reservation notification mask" feature command to mask reservation log.

- Add all the registrants that need to be freed to a temporary list fist, and then after calling synchronize_rcu(), release all the registrants on the temporary list.

- Fix the resv log page is random when there is no resv log page.

- Change nvmet_is_host_still_connected() to nvmet_is_host_connected().

- Remove nvmet_pr_set_rtype_and_holder() and change nvmet_pr_create_new_resv() to nvmet_pr_create_new_reservation().

- Change nvmet_pr_find_registrant_by_hostid() to nvmet_pr_find_registrant().

- Change nvmet_pr_send_resv_released() to nvmet_pr_resv_released().

- Change __nvmet_pr_unregister_one() to nvmet_pr_unregister_one().

- In nvmet_pr_unreg_by_prkey(), nvmet_pr_unreg_by_prkey_except_hostid() and nvmet_pr_unreg_except_hostid(), first do unregistering and then do event sending.

Changes from v5 to v6:
- Use synchronize_rcu() and kfree() to free registrant instead of kfree_rcu().

- Remove nvmet_pr_register_check_rkey(), put the check into pr_lock warp. And refactor the nvmet_pr_register().

- Add the print fmt to the head.

- Add lockdep_is_held(&pr->pr_lock) condition to list_for_each_entry_rcu.

- Fix the bug in nvmet_pr_update_reg_attr(), when the change_attr hook return fail, we should not replace the holder.

Changes from v4 to v5:

- Use rculist macros to handle registration_list instead of list macros regardless of in mutex lock or not.

- Use goto statement instead of return in nvmet_is_host_still_connected and __nvmet_pr_unregister_one.

- Add lockdep_assert_held and rcu_read_lock_held assert to many functions, if it's necessary.

- Add a comment to nvmet_execute_get_log_page_resv to explain how lost_count works.

- In nvmet_pr_clear, we should set holder to NULL first, I fixed this.

- Unify nvmet_pr_update_holder_rtype and __nvmet_pr_do_replace to nvmet_pr_update_reg_attr.

- Fix wrong nr_pages in nvmet_execute_get_log_page_resv.

- Fix the deadlock issue of nvmet_pr_exit_ns, put it out of the subsys lock.

Changes from v3 to v4:
- Use kfifo to handle resv log page instead of list, and also limit the resv log queue to 64.

- Change the function calling alignment style to:

    nvmet_pr_send_event_by_hostid(pr, hostid,
            NVME_PR_LOG_RESERVATOPM_PREEMPTED); 

- Put kmalloc out of rcu_read_lock in nvmet_execute_pr_report().

- Remove the goto in __nvmet_pr_unregister_one().

- Change generation to atomic_t, and remove nvmet_pr_inc_generation().

- In addtion, the number2 patch "nvmet: unify aer type enum" is not relate with this patch, so I will send it separately.

Changes from v2 to v3:
- Use rcu instead of rwlock to make IO path run faster, and put the rtype into the struct nvmet_pr_registrant.

- Limit the resv_log_list to 128.

- Change generation to atomic64.

- Put register rkey check to a warpper.

- Change nr_avl_pages to nr_pages.

- Use NVME_SC_SUCCESS instead of 0.

- Change kmalloc param to let it not sleep in mutex lock.

Changes from v1 to v2:
- Implement the reservation notification report, includes registration preempted, reservation released and reservation preempted. And also handle the reservation log page available event and send get
reservation log page command to clear log page at host.

- Put the reservation check access after validate opcode. And remove opcodes which nvmet not implement yet check.
Now there is no admin opcode nvmet implemented needs reservation check,
so I dont add reservation check to admin command path.
Next we need to do reservation check includes the situation of nsid is
0xffffffff at each admin command path, if it is needed.

- Add reservation commands support in nvmet_get_cmd_effects_nvm().

- From Chaitanya, change the local variable tree style to make it cleaner, and add some comments about NVMe spec.
And also change others advice from chaitanya.

- Put the nvmet_pr_check_cmd_access and nvmet_parse_pr_cmd into reservation enable check warp.

- Remove kmem_cache instead to use kmalloc and kfree.

- Change others advice from Sagi.

- Add a blktest test case, this patch will be sent before these series of patches.

Guixin Liu (2):
nvme: add reservation command's defines
nvmet: support reservation feature

 drivers/nvme/target/Makefile      |    2 +-
 drivers/nvme/target/admin-cmd.c   |   24 +-
 drivers/nvme/target/configfs.c    |   27 +
 drivers/nvme/target/core.c        |   58 +-
 drivers/nvme/target/fabrics-cmd.c |    4 +-
 drivers/nvme/target/nvmet.h       |   58 +-
 drivers/nvme/target/pr.c          | 1186 +++++++++++++++++++++++++++++
 include/linux/nvme.h              |   68 ++
 8 files changed, 1415 insertions(+), 12 deletions(-)
 create mode 100644 drivers/nvme/target/pr.c
Signature · 1 line
2.43.0
author
#1 lore raw

PATCH v17 1/2 nvme: add reservation command's defines

This is a preparation patch for NVMeOF target reservation commands implantation.
Add the defines of reservation command, such as reservation log and sub operations.

---
 include/linux/nvme.h | 68 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 68 insertions(+)
include/linux/nvme.h +68 -0
index b58d9405d65e..44d048d68503 100644
@@ -2037,4 +2037,72 @@ struct nvme_completion {
 #define NVME_MINOR(ver)		(((ver) >> 8) & 0xff)
 #define NVME_TERTIARY(ver)	((ver) & 0xff)
 
+enum {
+	NVME_AEN_RESV_LOG_PAGE_AVALIABLE	= 0x00,
+};
+
+enum {
+	NVME_PR_LOG_EMPTY_LOG_PAGE			= 0x00,
+	NVME_PR_LOG_REGISTRATION_PREEMPTED		= 0x01,
+	NVME_PR_LOG_RESERVATION_RELEASED		= 0x02,
+	NVME_PR_LOG_RESERVATOPM_PREEMPTED		= 0x03,
+};
+
+enum {
+	NVME_PR_NOTIFY_BIT_REG_PREEMPTED		= 1,
+	NVME_PR_NOTIFY_BIT_RESV_RELEASED		= 2,
+	NVME_PR_NOTIFY_BIT_RESV_PREEMPTED		= 3,
+};
+
+struct nvme_pr_log {
+	__le64			count;
+	__u8			type;
+	__u8			nr_pages;
+	__u8			rsvd1[2];
+	__le32			nsid;
+	__u8			rsvd2[48];
+};
+
+struct nvmet_pr_register_data {
+	__le64	crkey;
+	__le64	nrkey;
+};
+
+struct nvmet_pr_acquire_data {
+	__le64	crkey;
+	__le64	prkey;
+};
+
+struct nvmet_pr_release_data {
+	__le64	crkey;
+};
+
+enum nvme_pr_capabilities {
+	NVME_PR_SUPPORT_PTPL				= 1,
+	NVME_PR_SUPPORT_WRITE_EXCLUSIVE			= 1 << 1,
+	NVME_PR_SUPPORT_EXCLUSIVE_ACCESS		= 1 << 2,
+	NVME_PR_SUPPORT_WRITE_EXCLUSIVE_REG_ONLY	= 1 << 3,
+	NVME_PR_SUPPORT_EXCLUSIVE_ACCESS_REG_ONLY	= 1 << 4,
+	NVME_PR_SUPPORT_WRITE_EXCLUSIVE_ALL_REGS	= 1 << 5,
+	NVME_PR_SUPPORT_EXCLUSIVE_ACCESS_ALL_REGS	= 1 << 6,
+	NVME_PR_SUPPORT_IEKEY_VER_1_3_DEF		= 1 << 7,
+};
+
+enum nvme_pr_register_action {
+	NVME_PR_REGISTER_ACT_REG		= 0,
+	NVME_PR_REGISTER_ACT_UNREG		= 1,
+	NVME_PR_REGISTER_ACT_REPLACE		= 1 << 1,
+};
+
+enum nvme_pr_acquire_action {
+	NVME_PR_ACQUIRE_ACT_ACQUIRE		= 0,
+	NVME_PR_ACQUIRE_ACT_PREEMPT		= 1,
+	NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT	= 1 << 1,
+};
+
+enum nvme_pr_release_action {
+	NVME_PR_RELEASE_ACT_RELEASE		= 0,
+	NVME_PR_RELEASE_ACT_CLEAR		= 1,
+};
+
 #endif /* _LINUX_NVME_H */
Signature · 1 line
2.43.0
author
#2 lore raw

PATCH v17 2/2 nvmet: support reservation feature

In reply to Guixin Liu

This patch implements the reservation feature, including:
1. reservation register(register, unregister and replace).
2. reservation acquire(acquire, preempt, preempt and abort).
3. reservation release(release and clear).
4. reservation report.
5. set feature and get feature of reservation notify mask.
6. get log page of reservation event.

Not supported:
1. persistent reservation through power loss.

Test cases:
Use nvme-cli and fio to test all implemented sub features:
1. use nvme resv-register to register host a registrant or unregister or replace a new key.
2. use nvme resv-acquire to set host to the holder, and use fio to send read and write io in all reservation type. And also test preempt and "preempt and abort".
3. use nvme resv-report to show all registrants and reservation status.
4. use nvme resv-release to release all registrants.
5. use nvme get-log to get events generated by the preceding operations.

In addition, make reservation configurable, one can set ns to support reservation before enable ns. The default of resv_enable is false.

---
 drivers/nvme/target/Makefile      |    2 +-
 drivers/nvme/target/admin-cmd.c   |   24 +-
 drivers/nvme/target/configfs.c    |   27 +
 drivers/nvme/target/core.c        |   58 +-
 drivers/nvme/target/fabrics-cmd.c |    4 +-
 drivers/nvme/target/nvmet.h       |   58 +-
 drivers/nvme/target/pr.c          | 1186 +++++++++++++++++++++++++++++
 7 files changed, 1347 insertions(+), 12 deletions(-)
 create mode 100644 drivers/nvme/target/pr.c
drivers/nvme/target/Makefile +1 -1
index c402c44350b2..f2b025bbe10c 100644
@@ -10,7 +10,7 @@ obj-$(CONFIG_NVME_TARGET_FCLOOP)	+= nvme-fcloop.o
 obj-$(CONFIG_NVME_TARGET_TCP)		+= nvmet-tcp.o
 
 nvmet-y		+= core.o configfs.o admin-cmd.o fabrics-cmd.o \
-			discovery.o io-cmd-file.o io-cmd-bdev.o
+			discovery.o io-cmd-file.o io-cmd-bdev.o pr.o
 nvmet-$(CONFIG_NVME_TARGET_DEBUGFS)	+= debugfs.o
 nvmet-$(CONFIG_NVME_TARGET_PASSTHRU)	+= passthru.o
 nvmet-$(CONFIG_BLK_DEV_ZONED)		+= zns.o
drivers/nvme/target/admin-cmd.c +23 -1
index 081f0473cd9e..19428745c795 100644
@@ -176,6 +176,10 @@ static void nvmet_get_cmd_effects_nvm(struct nvme_effects_log *log)
 	log->iocs[nvme_cmd_read] =
 	log->iocs[nvme_cmd_flush] =
 	log->iocs[nvme_cmd_dsm]	=
+	log->iocs[nvme_cmd_resv_acquire] =
+	log->iocs[nvme_cmd_resv_register] =
+	log->iocs[nvme_cmd_resv_release] =
+	log->iocs[nvme_cmd_resv_report] =
 		cpu_to_le32(NVME_CMD_EFFECTS_CSUPP);
 	log->iocs[nvme_cmd_write] =
 	log->iocs[nvme_cmd_write_zeroes] =
@@ -340,6 +344,8 @@ static void nvmet_execute_get_log_page(struct nvmet_req *req)
 		return nvmet_execute_get_log_cmd_effects_ns(req);
 	case NVME_LOG_ANA:
 		return nvmet_execute_get_log_page_ana(req);
+	case NVME_LOG_RESERVATION:
+		return nvmet_execute_get_log_page_resv(req);
 	}
 	pr_debug("unhandled lid %d on qid %d\n",
 	       req->cmd->get_log_page.lid, req->sq->qid);
@@ -433,7 +439,8 @@ static void nvmet_execute_identify_ctrl(struct nvmet_req *req)
 	id->nn = cpu_to_le32(NVMET_MAX_NAMESPACES);
 	id->mnan = cpu_to_le32(NVMET_MAX_NAMESPACES);
 	id->oncs = cpu_to_le16(NVME_CTRL_ONCS_DSM |
-			NVME_CTRL_ONCS_WRITE_ZEROES);
+			NVME_CTRL_ONCS_WRITE_ZEROES |
+			NVME_CTRL_ONCS_RESERVATIONS);
 
 	/* XXX: don't report vwc if the underlying device is write through */
 	id->vwc = NVME_CTRL_VWC_PRESENT;
@@ -551,6 +558,15 @@ static void nvmet_execute_identify_ns(struct nvmet_req *req)
 	id->nmic = NVME_NS_NMIC_SHARED;
 	id->anagrpid = cpu_to_le32(req->ns->anagrpid);
 
+	if (req->ns->pr.enable)
+		id->rescap = NVME_PR_SUPPORT_WRITE_EXCLUSIVE |
+			NVME_PR_SUPPORT_EXCLUSIVE_ACCESS |
+			NVME_PR_SUPPORT_WRITE_EXCLUSIVE_REG_ONLY |
+			NVME_PR_SUPPORT_EXCLUSIVE_ACCESS_REG_ONLY |
+			NVME_PR_SUPPORT_WRITE_EXCLUSIVE_ALL_REGS |
+			NVME_PR_SUPPORT_EXCLUSIVE_ACCESS_ALL_REGS |
+			NVME_PR_SUPPORT_IEKEY_VER_1_3_DEF;
+
 	memcpy(&id->nguid, &req->ns->nguid, sizeof(id->nguid));
 
 	id->lbaf[0].ds = req->ns->blksize_shift;
@@ -861,6 +877,9 @@ void nvmet_execute_set_features(struct nvmet_req *req)
 	case NVME_FEAT_WRITE_PROTECT:
 		status = nvmet_set_feat_write_protect(req);
 		break;
+	case NVME_FEAT_RESV_MASK:
+		status = nvmet_set_feat_resv_notif_mask(req, cdw11);
+		break;
 	default:
 		req->error_loc = offsetof(struct nvme_common_command, cdw10);
 		status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
@@ -959,6 +978,9 @@ void nvmet_execute_get_features(struct nvmet_req *req)
 	case NVME_FEAT_WRITE_PROTECT:
 		status = nvmet_get_feat_write_protect(req);
 		break;
+	case NVME_FEAT_RESV_MASK:
+		status = nvmet_get_feat_resv_notif_mask(req);
+		break;
 	default:
 		req->error_loc =
 			offsetof(struct nvme_common_command, cdw10);
drivers/nvme/target/configfs.c +27 -0
index 685e89b35d33..eeee9e9b854c 100644
@@ -769,6 +769,32 @@ static ssize_t nvmet_ns_revalidate_size_store(struct config_item *item,
 
 CONFIGFS_ATTR_WO(nvmet_ns_, revalidate_size);
 
+static ssize_t nvmet_ns_resv_enable_show(struct config_item *item, char *page)
+{
+	return sysfs_emit(page, "%d\n", to_nvmet_ns(item)->pr.enable);
+}
+
+static ssize_t nvmet_ns_resv_enable_store(struct config_item *item,
+		const char *page, size_t count)
+{
+	struct nvmet_ns *ns = to_nvmet_ns(item);
+	bool val;
+
+	if (kstrtobool(page, &val))
+		return -EINVAL;
+
+	mutex_lock(&ns->subsys->lock);
+	if (ns->enabled) {
+		pr_err("the ns:%d is already enabled.\n", ns->nsid);
+		mutex_unlock(&ns->subsys->lock);
+		return -EINVAL;
+	}
+	ns->pr.enable = val;
+	mutex_unlock(&ns->subsys->lock);
+	return count;
+}
+CONFIGFS_ATTR(nvmet_ns_, resv_enable);
+
 static struct configfs_attribute *nvmet_ns_attrs[] = {
 	&nvmet_ns_attr_device_path,
 	&nvmet_ns_attr_device_nguid,
@@ -777,6 +803,7 @@ static struct configfs_attribute *nvmet_ns_attrs[] = {
 	&nvmet_ns_attr_enable,
 	&nvmet_ns_attr_buffered_io,
 	&nvmet_ns_attr_revalidate_size,
+	&nvmet_ns_attr_resv_enable,
 #ifdef CONFIG_PCI_P2PDMA
 	&nvmet_ns_attr_p2pmem,
 #endif
drivers/nvme/target/core.c +52 -6

137 lines not on this page · raw message

drivers/nvme/target/fabrics-cmd.c +1 -3
index c4b2eddd5666..28a84af1b4c0 100644
@@ -245,12 +245,10 @@ static void nvmet_execute_admin_connect(struct nvmet_req *req)
 	d->subsysnqn[NVMF_NQN_FIELD_LEN - 1] = '\0';
 	d->hostnqn[NVMF_NQN_FIELD_LEN - 1] = '\0';
 	status = nvmet_alloc_ctrl(d->subsysnqn, d->hostnqn, req,
-				  le32_to_cpu(c->kato), &ctrl);
+				  le32_to_cpu(c->kato), &ctrl, &d->hostid);
 	if (status)
 		goto out;
 
-	uuid_copy(&ctrl->hostid, &d->hostid);
-
 	dhchap_status = nvmet_setup_auth(ctrl);
 	if (dhchap_status) {
 		pr_err("Failed to setup authentication, dhchap status %u\n",
drivers/nvme/target/nvmet.h +57 -1

116 lines not on this page · raw message

drivers/nvme/target/pr.c new file +1186 -0

1,186 lines not on this page · raw message

Signature · 1 line
2.43.0
maintainer
#3 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

drivers/nvme/target/pr.c line 8

Quoted 13 lines · Guixin Liu

new file mode 100644
index 000000000000..aa69656a0ac5

drivers/nvme/target/pr.c new file +8 -0
@@ -0,0 +1,1186 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * NVMe over Fabrics Persist Reservation.
+ * Copyright (c) 2024 Guixin Liu, Alibaba Group.
+ * All rights reserved.
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/unaligned.h>

I don't find any such file. Should this must have said <asm/unaligned.h> instead? Curious to know if the target side was actually tested.

author
#4 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 16 lines · Keith Busch

drivers/nvme/target/pr.c line 8

Quoted 13 lines · Guixin Liu

new file mode 100644
index 000000000000..aa69656a0ac5

drivers/nvme/target/pr.c new file +8 -0
@@ -0,0 +1,1186 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * NVMe over Fabrics Persist Reservation.
+ * Copyright (c) 2024 Guixin Liu, Alibaba Group.
+ * All rights reserved.
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/unaligned.h>

I don't find any such file. Should this must have said <asm/unaligned.h> instead? Curious to know if the target side was actually tested.

New 6.12-rc2 add a commit 5f60d5f6bbc12e782fac78110b0ee62698f3b576("move asm/unaligned.h to linux/unaligned.h"), I change my code to fit this.

You can find the discission in v14.

Best Regards,

Guixin Liu

maintainer
#5 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 19 lines · Guixin Liu
Quoted 16 lines · Keith Busch

drivers/nvme/target/pr.c line 8

Quoted 13 lines · Guixin Liu

new file mode 100644
index 000000000000..aa69656a0ac5

drivers/nvme/target/pr.c new file +8 -0
@@ -0,0 +1,1186 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * NVMe over Fabrics Persist Reservation.
+ * Copyright (c) 2024 Guixin Liu, Alibaba Group.
+ * All rights reserved.
+ */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/unaligned.h>

I don't find any such file. Should this must have said <asm/unaligned.h> instead? Curious to know if the target side was actually tested.

Quoted 4 lines

New 6.12-rc2 add a commit 5f60d5f6bbc12e782fac78110b0ee62698f3b576("move asm/unaligned.h to linux/unaligned.h"), I change my code to fit this.

You can find the discission in v14.

I see, thank you for the clarification. I hadn't been paying close enough attention at that discussion.

So, minor problem here. This new feature is 6.13 material, but our upstream block branch has forked before that commit. IMO, I feel introducing that header movement after the merge window is poorly timed...

Anyway, I don't think I can get away with something this large in 6.12, so it's going to have to be in our second 6.13 nvme pull request after the next rebase. I've set myself a reminder to merge this when that happens.

Thanks!

#6 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 8 lines · Keith Busch

I see, thank you for the clarification. I hadn't been paying close enough attention at that discussion.

So, minor problem here. This new feature is 6.13 material, but our upstream block branch has forked before that commit. IMO, I feel introducing that header movement after the merge window is poorly timed...

Quoted 4 lines

Anyway, I don't think I can get away with something this large in 6.12, so it's going to have to be in our second 6.13 nvme pull request after the next rebase. I've set myself a reminder to merge this when that happens.

This absolutely is not 6.12 material obviously. But having the block tree branched of before -rc2 seems a bit unusual and potentially problematic anyway. Jens, can you rebase or full a recent rc in?

maintainer
#7 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 9 lines · Christoph Hellwig
Quoted 8 lines · Keith Busch

I see, thank you for the clarification. I hadn't been paying close enough attention at that discussion.

So, minor problem here. This new feature is 6.13 material, but our upstream block branch has forked before that commit. IMO, I feel introducing that header movement after the merge window is poorly timed...

Quoted 8 lines
Quoted 4 lines

Anyway, I don't think I can get away with something this large in 6.12, so it's going to have to be in our second 6.13 nvme pull request after the next rebase. I've set myself a reminder to merge this when that happens.

This absolutely is not 6.12 material obviously. But having the block tree branched of before -rc2 seems a bit unusual and potentially problematic anyway. Jens, can you rebase or full a recent rc in?

I can rebase it, no pull in there yet.

Signature · 1 line
Jens Axboe
maintainer
#8 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 15 lines · Jens Axboe
Quoted 14 lines · Christoph Hellwig
Quoted 12 lines · Keith Busch

I see, thank you for the clarification. I hadn't been paying close enough attention at that discussion.

So, minor problem here. This new feature is 6.13 material, but our upstream block branch has forked before that commit. IMO, I feel introducing that header movement after the merge window is poorly timed...

Anyway, I don't think I can get away with something this large in 6.12, so it's going to have to be in our second 6.13 nvme pull request after the next rebase. I've set myself a reminder to merge this when that happens.

Quoted 5 lines
Quoted 3 lines

This absolutely is not 6.12 material obviously. But having the block tree branched of before -rc2 seems a bit unusual and potentially problematic anyway. Jens, can you rebase or full a recent rc in?

I can rebase it, no pull in there yet.

It's on 6.12-rc4 now.

Signature · 1 line
Jens Axboe
#9 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 1 line · Jens Axboe

It's on 6.12-rc4 now.

Thanks a lot!

maintainer
#10 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

In reply to Jens Axboe
Quoted 1 line · Jens Axboe

It's on 6.12-rc4 now.

And nvme-6.13 is now rebased to that as well.

maintainer
#11 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

In reply to Guixin Liu
Quoted 39 lines
drivers/nvme/target/pr.c new file +39 -0
+static void nvmet_execute_pr_acquire(struct nvmet_req *req)
+{
+	u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
+	bool ignore_key = nvmet_pr_parse_ignore_key(cdw10);
+	/* Reservation type, bit 15:08 */
+	u8 rtype = (u8)((cdw10 >> 8) & 0xff);
+	/* Reservation acquire action, bit 02:00 */
+	u8 acquire_act = cdw10 & 0x07;
+	struct nvmet_ctrl *ctrl = req->sq->ctrl;
+	struct nvmet_pr_acquire_data *d = NULL;
+	struct nvmet_pr *pr = &req->ns->pr;
+	struct nvmet_pr_registrant *reg;
+	u16 status = NVME_SC_SUCCESS;
+
+	if (ignore_key ||
+	    rtype < NVME_PR_WRITE_EXCLUSIVE ||
+	    rtype > NVME_PR_EXCLUSIVE_ACCESS_ALL_REGS) {
+		status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
+		goto out;
+	}
+
+	d = kmalloc(sizeof(*d), GFP_KERNEL);
+	if (!d) {
+		status = NVME_SC_INTERNAL;
+		goto out;
+	}
+
+	status = nvmet_copy_from_sgl(req, 0, d, sizeof(*d));
+	if (status)
+		goto free_data;
+
+	status = NVME_SC_RESERVATION_CONFLICT | NVME_STATUS_DNR;
+	mutex_lock(&pr->pr_lock);
+	list_for_each_entry_rcu(reg, &pr->registrant_list, entry,
+				lockdep_is_held(&pr->pr_lock)) {
+		if (uuid_equal(&reg->hostid, &ctrl->hostid) &&
+		    reg->rkey == le64_to_cpu(d->crkey)) {
+			status = __nvmet_execute_pr_acquire(req, reg,
+					acquire_act, rtype, d);

drivers/nvme/target/pr.c line 714

Quoted 10 lines
drivers/nvme/target/pr.c new file +10 -0
+			break;
+		}
+	}
+
+	if (!status && acquire_act == NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT) {
+		kfree(d);
+		INIT_WORK(&req->r.abort_work, nvmet_pr_do_abort);
+		queue_work(nvmet_wq, &req->r.abort_work);
+		return;
+	}

Is there a reason why you queue this here and not inside __nvmet_execute_pr_acquire like before?

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+	mutex_unlock(&pr->pr_lock);

Hmm... you keep this mutex taken and release it from the work element async... Not a great practice...

Any easy way you see to avoid this?

#12 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 54 lines · Sagi Grimberg
Quoted 39 lines
drivers/nvme/target/pr.c new file +39 -0
+static void nvmet_execute_pr_acquire(struct nvmet_req *req)
+{
+	u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
+	bool ignore_key = nvmet_pr_parse_ignore_key(cdw10);
+	/* Reservation type, bit 15:08 */
+	u8 rtype = (u8)((cdw10 >> 8) & 0xff);
+	/* Reservation acquire action, bit 02:00 */
+	u8 acquire_act = cdw10 & 0x07;
+	struct nvmet_ctrl *ctrl = req->sq->ctrl;
+	struct nvmet_pr_acquire_data *d = NULL;
+	struct nvmet_pr *pr = &req->ns->pr;
+	struct nvmet_pr_registrant *reg;
+	u16 status = NVME_SC_SUCCESS;
+
+	if (ignore_key ||
+	    rtype < NVME_PR_WRITE_EXCLUSIVE ||
+	    rtype > NVME_PR_EXCLUSIVE_ACCESS_ALL_REGS) {
+		status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
+		goto out;
+	}
+
+	d = kmalloc(sizeof(*d), GFP_KERNEL);
+	if (!d) {
+		status = NVME_SC_INTERNAL;
+		goto out;
+	}
+
+	status = nvmet_copy_from_sgl(req, 0, d, sizeof(*d));
+	if (status)
+		goto free_data;
+
+	status = NVME_SC_RESERVATION_CONFLICT | NVME_STATUS_DNR;
+	mutex_lock(&pr->pr_lock);
+	list_for_each_entry_rcu(reg, &pr->registrant_list, entry,
+				lockdep_is_held(&pr->pr_lock)) {
+		if (uuid_equal(&reg->hostid, &ctrl->hostid) &&
+		    reg->rkey == le64_to_cpu(d->crkey)) {
+			status = __nvmet_execute_pr_acquire(req, reg,
+					acquire_act, rtype, d);

drivers/nvme/target/pr.c line 714

Quoted 10 lines
drivers/nvme/target/pr.c new file +10 -0
+			break;
+		}
+	}
+
+	if (!status && acquire_act == NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT) {
+		kfree(d);
+		INIT_WORK(&req->r.abort_work, nvmet_pr_do_abort);
+		queue_work(nvmet_wq, &req->r.abort_work);
+		return;
+	}

Is there a reason why you queue this here and not inside __nvmet_execute_pr_acquire
like before?

Quoted 6 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+	mutex_unlock(&pr->pr_lock);

Hmm... you keep this mutex taken and release it from the work element async... Not a
great practice...

It's actually invalid for a mutex. You'd need a semaphore for that.

Guixin, can you test the code with lockdep enabled and see if there is other fallout as well? Thanks!

author
#13 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 59 lines · Christoph Hellwig
Quoted 53 lines · Sagi Grimberg
Quoted 39 lines
drivers/nvme/target/pr.c new file +39 -0
+static void nvmet_execute_pr_acquire(struct nvmet_req *req)
+{
+	u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
+	bool ignore_key = nvmet_pr_parse_ignore_key(cdw10);
+	/* Reservation type, bit 15:08 */
+	u8 rtype = (u8)((cdw10 >> 8) & 0xff);
+	/* Reservation acquire action, bit 02:00 */
+	u8 acquire_act = cdw10 & 0x07;
+	struct nvmet_ctrl *ctrl = req->sq->ctrl;
+	struct nvmet_pr_acquire_data *d = NULL;
+	struct nvmet_pr *pr = &req->ns->pr;
+	struct nvmet_pr_registrant *reg;
+	u16 status = NVME_SC_SUCCESS;
+
+	if (ignore_key ||
+	    rtype < NVME_PR_WRITE_EXCLUSIVE ||
+	    rtype > NVME_PR_EXCLUSIVE_ACCESS_ALL_REGS) {
+		status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
+		goto out;
+	}
+
+	d = kmalloc(sizeof(*d), GFP_KERNEL);
+	if (!d) {
+		status = NVME_SC_INTERNAL;
+		goto out;
+	}
+
+	status = nvmet_copy_from_sgl(req, 0, d, sizeof(*d));
+	if (status)
+		goto free_data;
+
+	status = NVME_SC_RESERVATION_CONFLICT | NVME_STATUS_DNR;
+	mutex_lock(&pr->pr_lock);
+	list_for_each_entry_rcu(reg, &pr->registrant_list, entry,
+				lockdep_is_held(&pr->pr_lock)) {
+		if (uuid_equal(&reg->hostid, &ctrl->hostid) &&
+		    reg->rkey == le64_to_cpu(d->crkey)) {
+			status = __nvmet_execute_pr_acquire(req, reg,
+					acquire_act, rtype, d);

drivers/nvme/target/pr.c line 714

Quoted 10 lines
drivers/nvme/target/pr.c new file +10 -0
+			break;
+		}
+	}
+
+	if (!status && acquire_act == NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT) {
+		kfree(d);
+		INIT_WORK(&req->r.abort_work, nvmet_pr_do_abort);
+		queue_work(nvmet_wq, &req->r.abort_work);
+		return;
+	}

Is there a reason why you queue this here and not inside __nvmet_execute_pr_acquire
like before?

Quoted 5 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+
+	mutex_unlock(&pr->pr_lock);

Hmm... you keep this mutex taken and release it from the work element async... Not a
great practice...

Quoted 4 lines

It's actually invalid for a mutex. You'd need a semaphore for that.

Guixin, can you test the code with lockdep enabled and see if there is other fallout as well? Thanks!

You mean use semaphore to replace mutex pr_lock? If this works, I

can send a v18.

Best Regards,

Guixin Liu

#14 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 6 lines · Guixin Liu
Quoted 2 lines

Guixin, can you test the code with lockdep enabled and see if there is other fallout as well? Thanks!

You mean use semaphore to replace mutex pr_lock? If this works, I

can send a v18.

That should work. Although semaphores are kinda frowned upon, so better write up a comment why they are needed as well.

author
#15 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 8 lines · Christoph Hellwig
Quoted 5 lines · Guixin Liu
Quoted 2 lines

Guixin, can you test the code with lockdep enabled and see if there is other fallout as well? Thanks!

You mean use semaphore to replace mutex pr_lock? If this works, I

can send a v18.

That should work. Although semaphores are kinda frowned upon, so better write up a comment why they are needed as well.

OK, the v18 will be soon.

Best Regards,

Guixin Liu

author
#16 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

In reply to Sagi Grimberg
Quoted 45 lines · Sagi Grimberg
Quoted 45 lines
drivers/nvme/target/pr.c new file +44 -0
+static void nvmet_execute_pr_acquire(struct nvmet_req *req)
+{
+    u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
+    bool ignore_key = nvmet_pr_parse_ignore_key(cdw10);
+    /* Reservation type, bit 15:08 */
+    u8 rtype = (u8)((cdw10 >> 8) & 0xff);
+    /* Reservation acquire action, bit 02:00 */
+    u8 acquire_act = cdw10 & 0x07;
+    struct nvmet_ctrl *ctrl = req->sq->ctrl;
+    struct nvmet_pr_acquire_data *d = NULL;
+    struct nvmet_pr *pr = &req->ns->pr;
+    struct nvmet_pr_registrant *reg;
+    u16 status = NVME_SC_SUCCESS;
+
+    if (ignore_key ||
+        rtype < NVME_PR_WRITE_EXCLUSIVE ||
+        rtype > NVME_PR_EXCLUSIVE_ACCESS_ALL_REGS) {
+        status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
+        goto out;
+    }
+
+    d = kmalloc(sizeof(*d), GFP_KERNEL);
+    if (!d) {
+        status = NVME_SC_INTERNAL;
+        goto out;
+    }
+
+    status = nvmet_copy_from_sgl(req, 0, d, sizeof(*d));
+    if (status)
+        goto free_data;
+
+    status = NVME_SC_RESERVATION_CONFLICT | NVME_STATUS_DNR;
+    mutex_lock(&pr->pr_lock);
+    list_for_each_entry_rcu(reg, &pr->registrant_list, entry,
+                lockdep_is_held(&pr->pr_lock)) {
+        if (uuid_equal(&reg->hostid, &ctrl->hostid) &&
+            reg->rkey == le64_to_cpu(d->crkey)) {
+            status = __nvmet_execute_pr_acquire(req, reg,
+                    acquire_act, rtype, d);
+            break;
+        }
+    }
+
+    if (!status && acquire_act == 

NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT) {

Quoted 9 lines
Quoted 5 lines

+        kfree(d);
+        INIT_WORK(&req->r.abort_work, nvmet_pr_do_abort); +        queue_work(nvmet_wq, &req->r.abort_work); +        return;
+    }

Is there a reason why you queue this here and not inside __nvmet_execute_pr_acquire like before?

Because I should add a "if" branch here to prevent unlock the pr_lock and complete

the request. Instead of twice "if", I merge them to one, just let __nvmet_execute_pr_acquire()

return status.

Quoted 7 lines
Quoted 2 lines

+
+    mutex_unlock(&pr->pr_lock);

Hmm... you keep this mutex taken and release it from the work element async... Not a great practice...

Any easy way you see to avoid this?

Sure, this is not a greate practice, I will think deeply and find another way.

Best Regards,

Guixin Liu

#17 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 67 lines · Guixin Liu
Quoted 46 lines · Sagi Grimberg
Quoted 46 lines
drivers/nvme/target/pr.c new file +44 -0
+static void nvmet_execute_pr_acquire(struct nvmet_req *req)
+{
+    u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
+    bool ignore_key = nvmet_pr_parse_ignore_key(cdw10);
+    /* Reservation type, bit 15:08 */
+    u8 rtype = (u8)((cdw10 >> 8) & 0xff);
+    /* Reservation acquire action, bit 02:00 */
+    u8 acquire_act = cdw10 & 0x07;
+    struct nvmet_ctrl *ctrl = req->sq->ctrl;
+    struct nvmet_pr_acquire_data *d = NULL;
+    struct nvmet_pr *pr = &req->ns->pr;
+    struct nvmet_pr_registrant *reg;
+    u16 status = NVME_SC_SUCCESS;
+
+    if (ignore_key ||
+        rtype < NVME_PR_WRITE_EXCLUSIVE ||
+        rtype > NVME_PR_EXCLUSIVE_ACCESS_ALL_REGS) {
+        status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
+        goto out;
+    }
+
+    d = kmalloc(sizeof(*d), GFP_KERNEL);
+    if (!d) {
+        status = NVME_SC_INTERNAL;
+        goto out;
+    }
+
+    status = nvmet_copy_from_sgl(req, 0, d, sizeof(*d));
+    if (status)
+        goto free_data;
+
+    status = NVME_SC_RESERVATION_CONFLICT | NVME_STATUS_DNR;
+    mutex_lock(&pr->pr_lock);
+    list_for_each_entry_rcu(reg, &pr->registrant_list, entry,
+                lockdep_is_held(&pr->pr_lock)) {
+        if (uuid_equal(&reg->hostid, &ctrl->hostid) &&
+            reg->rkey == le64_to_cpu(d->crkey)) {
+            status = __nvmet_execute_pr_acquire(req, reg,
+                    acquire_act, rtype, d);
+            break;
+        }
+    }
+
+    if (!status && acquire_act ==

NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT) {
+ kfree(d);

Quoted 9 lines

drivers/nvme/target/pr.c line 714

Quoted 4 lines
drivers/nvme/target/pr.c new file +4 -0
+        INIT_WORK(&req->r.abort_work, nvmet_pr_do_abort);
+        queue_work(nvmet_wq, &req->r.abort_work);
+        return;
+    }

Is there a reason why you queue this here and not inside __nvmet_execute_pr_acquire
like before?

Because I should add a "if" branch here to prevent unlock the pr_lock and complete

the request. Instead of twice "if", I merge them to one, just let __nvmet_execute_pr_acquire()

return status.

Quoted 3 lines
Quoted 2 lines
(quoted diff) +2 -0
+
+    mutex_unlock(&pr->pr_lock);
Quoted 9 lines
Quoted 5 lines

Hmm... you keep this mutex taken and release it from the work element async... Not a
great practice...

Any easy way you see to avoid this?

Sure, this is not a greate practice, I will think deeply and find another way.

Guixin, please, take a look whether keeping that mutex locked during aborting is actually needed or not. Probably there is no such need - pr_lock protects ns.pr.registrants_list which is not touched in abort procedure.

BR,
Dmitry

author
#18 lore raw

Re: [PATCH v17 2/2] nvmet: support reservation feature

Quoted 74 lines · Dmitry Bogdanov
Quoted 65 lines · Guixin Liu
Quoted 45 lines · Sagi Grimberg
Quoted 45 lines
drivers/nvme/target/pr.c new file +44 -0
+static void nvmet_execute_pr_acquire(struct nvmet_req *req)
+{
+    u32 cdw10 = le32_to_cpu(req->cmd->common.cdw10);
+    bool ignore_key = nvmet_pr_parse_ignore_key(cdw10);
+    /* Reservation type, bit 15:08 */
+    u8 rtype = (u8)((cdw10 >> 8) & 0xff);
+    /* Reservation acquire action, bit 02:00 */
+    u8 acquire_act = cdw10 & 0x07;
+    struct nvmet_ctrl *ctrl = req->sq->ctrl;
+    struct nvmet_pr_acquire_data *d = NULL;
+    struct nvmet_pr *pr = &req->ns->pr;
+    struct nvmet_pr_registrant *reg;
+    u16 status = NVME_SC_SUCCESS;
+
+    if (ignore_key ||
+        rtype < NVME_PR_WRITE_EXCLUSIVE ||
+        rtype > NVME_PR_EXCLUSIVE_ACCESS_ALL_REGS) {
+        status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
+        goto out;
+    }
+
+    d = kmalloc(sizeof(*d), GFP_KERNEL);
+    if (!d) {
+        status = NVME_SC_INTERNAL;
+        goto out;
+    }
+
+    status = nvmet_copy_from_sgl(req, 0, d, sizeof(*d));
+    if (status)
+        goto free_data;
+
+    status = NVME_SC_RESERVATION_CONFLICT | NVME_STATUS_DNR;
+    mutex_lock(&pr->pr_lock);
+    list_for_each_entry_rcu(reg, &pr->registrant_list, entry,
+                lockdep_is_held(&pr->pr_lock)) {
+        if (uuid_equal(&reg->hostid, &ctrl->hostid) &&
+            reg->rkey == le64_to_cpu(d->crkey)) {
+            status = __nvmet_execute_pr_acquire(req, reg,
+                    acquire_act, rtype, d);
+            break;
+        }
+    }
+
+    if (!status && acquire_act ==

NVME_PR_ACQUIRE_ACT_PREEMPT_AND_ABORT) {

Quoted 9 lines

drivers/nvme/target/pr.c line 714

Quoted 5 lines
drivers/nvme/target/pr.c new file +5 -0
+        kfree(d);
+        INIT_WORK(&req->r.abort_work, nvmet_pr_do_abort);
+        queue_work(nvmet_wq, &req->r.abort_work);
+        return;
+    }

Is there a reason why you queue this here and not inside __nvmet_execute_pr_acquire
like before?

Because I should add a "if" branch here to prevent unlock the pr_lock and complete

the request. Instead of twice "if", I merge them to one, just let __nvmet_execute_pr_acquire()

return status.

Quoted 2 lines
Quoted 2 lines
(quoted diff) +2 -0
+
+    mutex_unlock(&pr->pr_lock);
Quoted 8 lines
Quoted 5 lines

Hmm... you keep this mutex taken and release it from the work element async... Not a
great practice...

Any easy way you see to avoid this?

Sure, this is not a greate practice, I will think deeply and find another way.

Quoted 7 lines

Guixin, please, take a look whether keeping that mutex locked during aborting is actually needed or not. Probably there is no such need - pr_lock protects ns.pr.registrants_list which is not touched in abort procedure.

BR,
Dmitry

Well, I should prevent there is no reinit_completion() between complete()

and wait_for_completion(), this may happen if there is another "registion"

and "preempt and abort" when doing abort, although this is a host's bug.

Best Regards,

Guixin Liu